> ## Documentation Index
> Fetch the complete documentation index at: https://docs.artnex.app/llms.txt
> Use this file to discover all available pages before exploring further.

# Sync Lips to Audio with AI Avatars

> Artnex Lip Sync animates a portrait photo or video to match any audio track — create talking avatars with nine AI models from ByteDance, Kling, and more.

Lip Sync takes a portrait photo (or short video) and an audio track, then animates the subject's face — including mouth, expressions, and subtle head movements — to match the audio. The result is a realistic talking avatar video you can use for presentations, social content, explainers, or creative projects.

## Navigate to Lip Sync

Go to **Tools > Video > Lip Sync** in the Artnex sidebar.

## How to create a lip-sync video

<Steps>
  <Step title="Upload a portrait image">
    Click **Add Image** in the control bar to upload a photo of the person you want to animate. For best results, see the [reference image guidelines](#reference-image-guidelines) below.

    <Note>
      Image files must be under 25 MB. JPEG and PNG are both supported.
    </Note>
  </Step>

  <Step title="Upload an audio file">
    Click **Add Audio** to upload the audio track the avatar should speak. Any common audio format is accepted (MP3, WAV, M4A, etc.).

    <Note>
      Audio files must be under 25 MB. The InfiniteTalk Fast model supports audio tracks up to 10 minutes long.
    </Note>
  </Step>

  <Step title="Choose a model">
    Click the model selector button in the control bar to open the model dialog. Choose a model based on your quality, speed, and budget requirements.
  </Step>

  <Step title="Write an optional scene prompt">
    The prompt field is optional. You can use it to guide the visual style or scene context — for example: *"Professional studio setting, natural lighting, calm expression."*
  </Step>

  <Step title="Select resolution">
    Use the resolution button to choose between 480p and 720p output.
  </Step>

  <Step title="Generate">
    Click **Create** to start generation. The credit cost for the selected model is shown on the button. Generation typically takes 1–5 minutes depending on the model and audio length.
  </Step>

  <Step title="Download your video">
    When complete, the lip-sync video appears in your **My Creations** gallery. Click the **Download** icon to save it.
  </Step>
</Steps>

## Available models

<CardGroup cols={2}>
  <Card title="Avatar OmniHuman" icon="user">
    Animates a portrait photo into a lifelike avatar video with natural motion. Powered by ByteDance.

    **15,000 credits** per video
  </Card>

  <Card title="Avatar OmniHuman 1.5" icon="user">
    Enhanced OmniHuman generation with improved facial expressions and smoother motion.

    **18,000 credits** per video
  </Card>

  <Card title="InfiniteTalk Fast" icon="zap">
    Fast audio-driven talking avatar generation. Supports audio tracks up to 10 minutes long.

    **19,000 credits** per video
  </Card>

  <Card title="InfiniteTalk" icon="microphone">
    High-quality infinite lip-sync. Produces more natural mouth movements than the fast variant.

    **20,000 credits** per video
  </Card>

  <Card title="Kling v1 Avatar Standard" icon="sparkles">
    Standard quality Kling AI avatar lip-sync for general use.

    **21,000 credits** per video
  </Card>

  <Card title="Kling v2 Avatar Standard" icon="sparkles">
    Next-generation Kling avatar with improved expression fidelity at standard quality.

    **25,000 credits** per video
  </Card>

  <Card title="HunyuanAvatar" icon="face-smile">
    High-fidelity audio-driven avatar with emotion control. Produces expressive, cinematic results.

    **25,000 credits** per video
  </Card>

  <Card title="Kling v1 Avatar Pro" icon="crown">
    Professional-quality Kling avatar for detailed, high-resolution results.

    **50,000 credits** per video
  </Card>

  <Card title="Kling v2 Avatar Pro" icon="crown">
    Kling's best avatar model. Highest fidelity lip-sync with superior expression realism.

    **55,000 credits** per video
  </Card>
</CardGroup>

## Reference image guidelines

The quality of your output depends heavily on the source image you provide. Follow these guidelines to get the best results:

<CardGroup cols={2}>
  <Card title="Do" icon="check">
    * Use a **front-facing portrait** with the face clearly visible
    * Choose a photo with **even, natural lighting** and no harsh shadows
    * Use an image where **only one person** is in the frame
    * Make sure the face is **sharp and in focus**
    * Use a **neutral or slight smile** expression for the most natural animation
  </Card>

  <Card title="Avoid" icon="x">
    * Heavily rotated or side-profile faces
    * Images with sunglasses, masks, or face coverings
    * Small faces in a large scene — crop the image to centre the face
    * Blurry, grainy, or low-resolution photos
    * Multiple people in the same frame
  </Card>
</CardGroup>

<Tip>
  A headshot-style photo with the face filling most of the frame produces the most convincing animation. Portrait mode shots from a smartphone work well.
</Tip>

<Warning>
  Only animate faces with the explicit permission of the person depicted. Review Artnex's acceptable use policy before generating lip-sync videos of real individuals.
</Warning>